1. Field of the Invention
This invention relates generally to the manufacturing of high performance semiconductor devices. More specifically, this invention relates to the manufacture and testing of high performance semiconductor devices. Even more specifically, this invention relates to the manufacture and testing of high performance semiconductor devices using an image archival system to search for related images.
2. Discussion of the Related Art
In order to remain competitive, a semiconductor manufacturer must continuously increase the performance of the semiconductor integrated circuits being manufactured and at the same time, reduce the cost of the semiconductor integrated circuits. Part of the increase in performance and the reduction in cost of the semiconductor integrated circuits is accomplished by shrinking the device dimensions and by increasing the number of circuits per unit area on an integrated circuit chip. Another part of reducing the cost of a semiconductor chip is to increase the yield. As is known in the semiconductor manufacturing art, the yield of chips (also known as die) from each wafer is not 100% because of defects during the manufacturing process. The number of good chips obtained from a wafer determines the yield. As can be appreciated, chips that must be discarded because of a defect or defects increases the cost of the remaining usable chips because the cost of manufacturing must be amortized over the remaining usable chips.
A single semiconductor chip requires numerous process steps during manufacturing. These steps include oxidation, etching, metallization and wet chemical cleaning. Typically these process steps involve placing the wafer on which the semiconductor chips are being manufactured into different tools during the manufacturing process. The optimization of each of these process steps requires an understanding of a variety of chemical reactions and physical processes in order to produce high performance, high yield circuits. The ability to view and characterize the surface and interface layers of a semiconductor chip in terms of their morphology, chemical composition and distribution is an invaluable aid to those involved in research and development, process, problem solving, and failure analysis of integrated circuits. A major part of the analysis process is to capture defects and to analyze them completely to determine what caused the defects and to eliminate the cause.
In order to be able to quickly resolve process or equipment issues in the manufacture of semiconductor products, a great deal of time, effort and money is being expended by semiconductor manufacturers to capture and classify silicon based defects. Once a defect is caught and properly described and classified, work can begin to resolve the cause of the defect and to eliminate the cause of the defect. The biggest problem that faced semiconductor manufacturers was the inability of human inspectors to uniformly classify all defects consistently and without error. This problem was solved by the development of Automatic Defect Classification (ADC) systems.
One system for automatically classifying defects consists of the following methodological sequence. Gather a defect image from a review station. View the defect image and assign values to elemental descriptor terms called predicates that are general descriptors such as roundness, brightness, color, hue, graininess, etc. Assign a classification code to the defect based upon the values of all the predicates. A typical ADC system can have 40 or more quantifiable qualities and properties that can be predicates. Each predicate can have a specified range of values and a typical predicate can have a value assigned to it between 1 and 256. A value of 1 indicates that none of the value is present and a value of 256 indicates that the quality represented by the predicate is ideal. For example, a straight line would have a value of 1 for the predicate indicating roundness, whereas a perfect circle would have a value of 256 for the same predicate.
The classification code for each defect is determined by the system from the combination of all the predicate values assigned to the defect. The goal of the ADC system is to be able to uniquely describe all the defect types, in such a manner that a single classification code can be assigned to a defect which has been differentiated from all other defect types. This is accomplished by a system administrator who trains an artificial intelligence system to recognize various combinations and permutations of the 40 or more predicates to assign the same classification code to the same type of defect. This would result in a highly significant statistical confidence in the probability that the defect, and all other defects of the same type or class, will always be assigned the same classification code by the ADC system. These predicate values from the ADC system are stored in a database.
The current trend for semiconductor manufacturers is for increasing reliance on large integrated databases to provide analytical insight and yield prediction. These huge databases consist of quantitative data collected at various process modules in the manufacturing sequence as well as images of various features, such as defects alignment marks, focus matrices, etc., which provide a pictorial reference to problems or critical features. Currently the image data storage space utilization is superceding the data storage space at a rate of up to an order of magnitude more of total used storage space. Typically, a current database would have 80 GB consisting of 60-70 GB of image based information and 1-20 GB of data based information with the trend moving toward higher non-data based storage with video and sound becoming important components.
In the typical database most data is stored relationally, such that one can search the entire database for other examples exhibiting the same or similar characteristics of various parameters. While this is true for data based information, that is, lot numbers, critical dimensions, thicknesses, dates, operators, machines etc., this is virtually uniformly untrue for image-based data. The image itself may be linked to a certain defect or location at a certain layer of a certain lot, which can be relationally addressed, but information stored in the image is not retrievable relationally. For example, one can not retrieve all images having a red foreground with a repetitive dense pattern containing a small black defect with a halo using current storage methods.
Images have been described for purposes of ADC by a means of descriptors, which are measures of the various component parts that make the cohesive whole of an image. A typical descriptor can be color, roundness, brightness, intensity, contrast, repetitiveness, density, etc. These descriptive elements are typically used for comparison purposes with certain thumbprint values that have been generated within the ADC schema for classification purposes, and once a classification is determined the generated values are discarded. The problem with this method is that the images cannot be searched.
According to the present invention, the foregoing and other objects and advantages are attained by a method of analyzing defect images in a semiconductor manufacturing process wherein descriptors are assigned to images of defects.
In accordance with an aspect of the invention a production lot of wafers through a manufacturing a manufacturing process, scanning a selected wafer for defects and sending location information to a defect management system. The images of the defects are analyzed in an analysis tool that assigns descriptors to the image. The defect management system stores images, the assigned descriptors and linkage data in a database.
In accordance with another aspect of the invention a selected image is reviewed at a review station, which assign descriptors to the selected image that describe a desired feature of the selected image that is to be analyzed.
In accordance with another aspect of the invention the database is searched for images having descriptors assigned to the selected image.
In accordance with still another aspect of the invention the review station assigns refined descriptors to the selected image and the database is searched for images having the refined descriptors assigned to the selected image.
The described method thus effectively provides a method of analyzing defect images in a semiconductor manufacturing process wherein descriptors are assigned to images of defects whereby the images can be searched in a relational database.
The present invention is better understood upon consideration of the detailed description below, in conjunction with the accompanying drawings. As will become readily apparent to those skilled in the art from the following description, there is shown and described an embodiment of this invention simply by way of illustration of the best mode to carry out the invention. As will be realized, the invention is capable of other embodiments and its several details are capable of modifications in various obvious aspects, all without departing from the scope of the invention. Accordingly, the drawings and detailed description will be regarded as illustrative in nature and not as restrictive.